linux-kernel - Re: [PATCH v2 0/7] [PATCH v2 0/7] Add support of OV9655 camera

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <2dd3402e-55b0-231d-878f-5ba95ee8cb36@st.com>
Date:   Tue, 18 Jul 2017 12:53:12 +0000
From:   Hugues FRUCHET <hugues.fruchet@...com>
To:     "H. Nikolaus Schaller" <hns@...delico.com>,
        Sylwester Nawrocki <snawrocki@...nel.org>,
        Hans Verkuil <hverkuil@...all.nl>
CC:     Guennadi Liakhovetski <g.liakhovetski@....de>,
        Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Maxime Coquelin <mcoquelin.stm32@...il.com>,
        Alexandre TORGUE <alexandre.torgue@...com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        devicetree <devicetree@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        "Yannick FERTRE" <yannick.fertre@...com>,
        Benjamin Gaignard <benjamin.gaignard@...aro.org>,
        linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
        "linux-media@...r.kernel.org" <linux-media@...r.kernel.org>
Subject: Re: [PATCH v2 0/7] [PATCH v2 0/7] Add support of OV9655 camera



On 07/18/2017 02:17 PM, H. Nikolaus Schaller wrote:
> Hi,
> 
>> Am 18.07.2017 um 13:59 schrieb Hans Verkuil <hverkuil@...all.nl>:
>>
>> On 12/07/17 22:01, Sylwester Nawrocki wrote:
>>> Hi Hugues,
>>>
>>> On 07/03/2017 11:16 AM, Hugues Fruchet wrote:
>>>> This patchset enables OV9655 camera support.
>>>>
>>>> OV9655 support has been tested using STM32F4DIS-CAM extension board
>>>> plugged on connector P1 of STM32F746G-DISCO board.
>>>> Due to lack of OV9650/52 hardware support, the modified related code
>>>> could not have been checked for non-regression.
>>>>
>>>> First patches upgrade current support of OV9650/52 to prepare then
>>>> introduction of OV9655 variant patch.
>>>> Because of OV9655 register set slightly different from OV9650/9652,
>>>> not all of the driver features are supported (controls). Supported
>>>> resolutions are limited to VGA, QVGA, QQVGA.
>>>> Supported format is limited to RGB565.
>>>> Controls are limited to color bar test pattern for test purpose.
>>>
>>> I appreciate your efforts towards making a common driver but IMO it would be
>>> better to create a separate driver for the OV9655 sensor.  The original driver
>>> is 1576 lines of code, your patch set adds half of that (816).  There are
>>> significant differences in the feature set of both sensors, there are
>>> differences in the register layout.  I would go for a separate driver, we
>>> would then have code easier to follow and wouldn't need to worry about possible
>>> regressions.  I'm afraid I have lost the camera module and won't be able
>>> to test the patch set against regressions.
>>>
>>> IMHO from maintenance POV it's better to make a separate driver. In the end
>>> of the day we wouldn't be adding much more code than it is being done now.
>>
>> I agree. We do not have great experiences in the past with trying to support
>> multiple variants in a single driver (unless the diffs are truly small).
> 
> Well,
> IMHO the diffs in ov965x are smaller (but untestable because nobody seems
> to have an ov9650/52 board) than within the bq27xxx chips, but I can dig out
> an old pdata based separate ov9655 driver and extend that to become DT compatible.
> 
> I had abandoned that separate approach in favour of extending the ov965x driver.
> 
> Have to discuss with Hugues how to proceed.
> 
> BR and thanks,
> Nikolaus
> 

As Sylwester and Hans, I'm also in flavour of a separate driver, the 
fact that register set seems similar but in fact is not and that we 
cannot test for non-regression of 9650/52 are killer for me to continue 
on a single driver.
We can now restart from a new fresh state of the art sensor driver 
getting rid of legacy (pdata, old gpio, etc...).

BR,
Hugues.